Professional Documents
Culture Documents
Bijlage 1 - Reciprocal Teaching and Teacher Skills - Compleet
Bijlage 1 - Reciprocal Teaching and Teacher Skills - Compleet
a
University of Twente, P.O. Box 217, 7500 AE Enschede, The Netherlands
roelvansteensel@gmail.com
c
Kohnstamm Institute, P.O. Box 94208, 1090 GE Amsterdam, The Netherlands
Avangelderen@kohnstamm.uva.nl
d
University of Twente, P.O. Box 217, 7500 AE Enschede, The Netherlands
p.j.c.sleegers@utwente.nl
RECIPROCAL TEACHING AND TEACHER SKILLS
Abstract
Low-achieving adolescents are known to have difficulties with reading comprehension. This
article discusses how reciprocal teaching can improve low-achieving adolescents’ reading
comprehension in natural classroom settings (as opposed to small-group settings) and to what
extent intervention effects are dependent on teacher behavior. Over the course of one year,
experimental teachers (n=10) were given extensive training and coaching aimed at using
principles of reciprocal teaching, while control teachers (n=10) used their regular teaching
modeling, and support of group work, and were performed in both experimental and control
classes, comprising a total of 369 students (grade 7). Our study shows that reciprocal teaching
teachers provided high-quality strategy instruction. In addition, results suggest that the
Organisation for Economic Co-operation and Development [OECD], 2003; OECD, 2014).
Since reading comprehension is a fundamental skill in many school subjects, difficulties can
have serious implications for students’ educational success and, consequently, for their later
societal careers. Evidence-based reading comprehension programs that target low achieving
adolescents are thus of vital importance. In this study, we analyze the effects of an
Reciprocal teaching
Reciprocal teaching (Palincsar & Brown, 1984) is a widely used method of instructing and
fading; and c) students practicing and discussing reading strategies with other students,
guided and coached by the teacher. Reciprocal teaching assumes a gradual shift of
responsibility for the learning process from teacher to student, which includes the teacher
explicitly modeling the use of reading strategies (Rosenhine & Meister, 1994) as well as
scaffolding the application of reading strategies within the groups of students working
RECIPROCAL TEACHING AND TEACHER SKILLS
increasingly more capable of regulating their own reading process. In this study, we consider
including direct instruction of reading strategies, teacher and student modeling, and group
work.
Many studies have confirmed the positive effects of reciprocal teaching (Rosenshine &
Meister, 1994; Kelly, Moore, & Tuck, 2001; Spörer, Brunstein, & Kieschke, 2009). In a
review by Rosenshine and Meister (1994), sixteen studies were analyzed. The authors found
an overall positive effect on reading comprehension, with a median Cohen’s effect size value
(d =.32) for standardized tests and .88 for researcher-developed tests. They also examined the
effects of several moderator variables, of which two are particularly relevant for the current
study: group size and type of interventionist (teacher or researcher). Regarding the former,
they found contradictory results for studies where reciprocal reaching was applied in large
groups (>18), with two studies showing positive significant results, one study with mixed
results, and one with nonsignificant results. Regarding the latter, they also found ambiguous
results for teacher-led interventions, with two studies with positive significant results, three
studies with mixed results and two studies with nonsignificant results. Thus, whether larger
group size or teacher-led reciprocal teaching matter in finding positive results, is undecided.
readers between Grades 6 and 9 (Edmonds et al., 2009), seven studies focusing on reading
comprehension were included. Most of these studies included some kind of instruction in
reading strategies, with two of them using reciprocal teaching. The overall Cohen’s effect
size (d =1.23) on reading comprehension was very large. However, effects of possible
RECIPROCAL TEACHING AND TEACHER SKILLS
moderators such as those reported by Rosenshine and Meister (1994) were not reported in
this synthesis, which makes it difficult to draw definite conclusions about the question
whether reciprocal teaching is effective in regular classroom settings with regular teachers.
Reciprocal teaching was originally designed by Palincsar and Brown (1987) for small-
group tutoring under the guidance of experts, in which small groups of students were taken
out of the classroom. In a whole-class setting, where 15-30 students are present, such
extensive guidance as is provided in a small group might be quite difficult, if at all possible,
as the teacher needs to pay attention to multiple groups of students within the classroom.
Furthermore, small-group settings are often used in controlled experiments where the
comparison to researchers, who have extensive background knowledge about the theoretical
basis of reciprocal teaching, the quality of implementation might be different for teachers,
Whereas reviews do not allow such conclusions on, studies in which teachers were
that the quality of implementation is indeed a serious problem (Duffy, 1993; Seymoor &
Osana, 2003; Hacker & Tenent, 2002). Duffy (1993) described the process of teachers
becoming experts in reading strategies. Teachers were followed during the implementation of
study, the teachers were interviewed several times. A major conclusion from this study is that
teachers realized that being able to model the use of strategies and explicitly relating strategy-
use to text is not enough to induce strategic thinking in students that is useful for integrating
Seymour and Osana (2003) found that teachers faced similar problems when they
were trained in reciprocal teaching. In their study, two teachers were trained and observed
RECIPROCAL TEACHING AND TEACHER SKILLS
during the implementation of reciprocal teaching. Interviews with the teachers revealed that
their knowledge about reading strategies increased substantially during training, but their
understanding of didactic principles was not developed optimally. Particularly, the teachers
still did not fully understand what scaffolding entails at the end of the training.
These findings are corroborated by Hacker and Tenent (2002), who studied the
application of reciprocal teaching in regular classrooms (Hacker & Tenent, 2002). They
examined the way 17 teachers implemented reciprocal teaching and adapted the method to
their own teaching practice over the course of three years. The researchers showed that
teachers found it difficult to maintain the original format. First, they found that “student
dialogues were hampered because of the students’ poor group discourse skills” as well as the
poor application of reading strategies by the students, resulting in the observation “that there
really was little for them to discuss” (Hacker & Tenent, 2002, p. 703). To deal with those
problems, the teachers extended whole-class instruction of reading strategies to at least two
months and they provided more scaffolding of strategy use in different kinds of contexts
while at the same time providing scaffolding of the collaborative process. In other words, the
approach, which hampered the implementation of collaborative group work in discussing and
practicing reading strategies. Second, Hacker and Tenent (2002) found that the students had
difficulties with using all four reading strategies (predicting, questioning, summarizing and
clarifying). Not all strategies were used, and the strategies that were used (summarizing and
questioning) were “often being used inadequately” (p. 702). Students tended to ask
superficial questions instead of making elaborations and reflections, and their strategy use
possible explanations of why previous experimental studies do not always support the success
RECIPROCAL TEACHING AND TEACHER SKILLS
requires not only expert knowledge about the use of reading strategies on the part of the
teachers, but also skills for regulating students’ collaborative process in different groups
simultaneously.
Our study aims to contribute to existing knowledge in two ways. First, we examined whether
prevocational education, in which regular teachers are delivering the lessons (Woolley,
2011). Second, we analyzed whether intervention effects were moderated by the extent to
which teachers were able to apply these principles. Issues of treatment fidelity have received
little attention in reading intervention research (National Reading Panel, 2000; Rosenhine &
Meister, 1994; Edmonds et al., 2009). Therefore, this study aims to add to the research base
reciprocal teaching. This allows insight into the conditions under which the treatment will be
teaching (strategy instruction, modeling and group work) moderate effects on reading
comprehension?
RECIPROCAL TEACHING AND TEACHER SKILLS
Method
Design
2002). The design included one independent variable (treatment vs control) and one
Gender was included, because generally girls are shown to have substantially greater reading
skill than boys (e.g., Logan & Johnston, 2009). Vocabulary knowledge and IQ were included,
as theoretical models suggest that reading comprehension draws heavily on both abilities
(e.g., Just & Carpenter, 1976, 2004; LaBerge & Samuels, 1974; Rumelhart, 2004; Samuels,
2004), an assumption that is confirmed by much empirical evidence (e.g., Ouelette & Beers,
2010; Van Gelderen et al., 2004, 2007; Verhoeven & Van Leeuwe, 2008). Finally, we
included three moderator variables, covering the three didactic principles behind our
treatment: direct instruction of reading strategies, teacher and student modeling, and group
work. Randomization was applied at the teacher/class level. At every participating school two
classes were selected, each with their own Dutch language teacher. These were then
In the timeline in Figure 1 a depiction is given of all the research activities that took
______________________________________________________________________
______________________________________________________________________
Our study focused on low achievers. Our operationalization of low achievement was based on
educational track. The Netherlands have a tracked system of secondary education. After
education, senior general secondary education, pre-university education—on the basis of their
scores on a general attainment test and their educational performance as assessed by their
primary school teachers (Ministry of Education, Culture, & Science, 2006). Since students in
prevocational education are generally characterized by poor reading skills (Dutch Education
Inspectorate, 2008; Gille, Loijens, Noijons, & Zwitser, 2010), we selected our sample from
We recruited schools in two ways. First, we contacted schools that had participated in
a previous study on low achieving readers. Second, we contacted schools via a digital
community of Dutch language teachers. Schools had to meet the following five criteria:
The teachers were prepared to take part in the randomization procedure, implying that
a) if their class was assigned to the treatment condition, they were prepared to take
part in our training and coaching program and to weekly give our experimental
lessons; and b) if their class was assigned to the control condition, they were prepared
to not use our program nor discuss its contents with the colleague in the treatment
condition.
Control teachers were requested to use their regular language program during the
language classes.
RECIPROCAL TEACHING AND TEACHER SKILLS
Ten different schools in different parts of the Netherlands were willing to participate. Within
each school, two teachers volunteered. Randomization was done at the class level within each
school, resulting in a total of ten experimental and ten control classes, each with their teacher,
divided over the ten schools. At the start of the study, these classes comprised 369 students,
of which 189 were in the treatment condition (51%) and 180 in the control condition (49%).
The students’ mean age was 13.01 years (SD = 0.52) at the start of the project. There was no
statistically significant difference between the two conditions on this variable (t(366) = -1.27,
p = .20). There were relatively more girls in the sample (n= 200; 54%) than boys (n= 169;
46%), with relatively more girls than boys (59 vs 41%) in the treatment condition. The
distribution in the control condition, however, was more equal (49 vs 51%). The difference in
distribution between the two conditions was statistically significant (χ2(1) = 3.99, p = .046).
More female than male teachers participated in the study (N = 15 vs. N = 5), with two
male teachers in the treatment group and three males in the control group. The mean age of
the teachers was 46.40 years (SD = 11.12). On average they had 13.50 (SD = 13.73, min = 1,
max = 38) years of teaching experience in secondary education. No differences were found
between the conditions on either variable, (t(14) = -.45, p = .66) and (t(14) = .053, p = .96),
respectively.
Treatment
Our intervention consisted of the training of teachers in the use of the three related
instructional strategies of reciprocal teaching (Palincsar & Brown, 1984), that is:
1. Direct instruction of research-based reading strategies (see further). For each strategy,
it was emphasized what the strategy entailed, how to use the strategy, when to use the
strategy and why to use the strategy (Veenman, 2006). Thus, teachers were required
RECIPROCAL TEACHING AND TEACHER SKILLS
procedural knowledge.
2. Teacher and student modeling. Teachers were trained to model the use of reading
strategies during plenary instruction by thinking aloud when reading text. They
encouraged students to take over this role, both plenary and in small group sessions.
3. Group work. The primary objective of encouraging students to work in groups was to
have them collaboratively apply reading strategies while thinking aloud during text
reading. Teachers were given instructions on how to give feedback to the groups of
students working together. For example, if a teacher noticed that the students were
struggling with the application of a reading strategy, the teacher was to model this
strategy again and encourage and aid the students in doing this themselves.
Students received weekly lessons over a period of seven months within one school year.
During the school year, the experimental teachers were trained and coached.
With respect to strategy instruction the intervention focused on five strategies that were
shown to be related to reading comprehension in previous research (Dole, Duffy, Roehler, &
Pearson, 1991; Palincsar & Brown, 1984; Pressley & Afflerbach, 1995):
1. Predicting. On the basis of text features such as title, subheadings, and pictures,
students are instructed to make predictions about text content before reading, and to
to focus on main ideas and ignore irrelevant details as well as to check their
3. Self-questioning. Students are instructed to generate questions about the text being
4. Clarifying. When confronted with a word or passage they do not understand, students
are instructed to reread, read ahead, or, in the case of an unknown word, analyze it,
and see whether its meaning can be inferred by looking at parts of the word.
5. Interpreting cohesive ties. Students are instructed to look for relationships between
sentences or paragraphs that are connected, e.g. by using ‘signal words’ (different
types of connectives).
The intervention was offered in the context of an existing program called “Nieuwsbegrip”®,
developed by the CED Group in Rotterdam (“Comprehension of news”, CED Group, 2011).
Lessons were developed weekly by a team of developers at the CED Group. They were based
on recent news texts (i.e., texts that had been issued the week before) about subjects close to
students’ everyday life (e.g., sugar in energy drinks, abdication of the Dutch queen, or 20
years of text messaging). The use of topical, interesting texts aimed to increase students’ task
motivation (Guthrie & Wigfield, 2000; Schiefele, 1999). The lessons could be downloaded
by teachers from the program website (www.nieuwsbegrip.nl) every week, starting Monday
evening.
Lessons were provided in sequences of six weeks. Each sequence consisted of six
weekly lessons (approximately 45 minutes per lesson). In each of the first five lessons, the
focus was on one reading strategy that was practiced in a central strategy assignment that was
provided on a work sheet. In addition, students could work on other assignments (i.e.,
Each of the five strategies was trained several times during the year. This cyclical
approach was assumed to result in the consolidation of strategy knowledge. In the final lesson
of each sequence all strategies were practiced simultaneously. The idea behind this was that
students have to be able to apply all strategies together during the reading process, selecting
RECIPROCAL TEACHING AND TEACHER SKILLS
the right strategy at the right moment. Figure 2 provides for each reading strategy an example
of an assignment.
____________________________________________________________________
____________________________________________________________________
Treatment teachers took part in an extensive training and coaching program that was
conducted by teacher trainers from the Rotterdam University of Applied Sciences, who had,
in turn, been trained by the first three authors. In the first phase (October 2011-January 2012),
teachers participated in three one-hour training sessions. In Session 1, they received general,
practical information about the program (e.g., how to use the program website), theoretical
information about the reading process and its components, and basic information about the
program’s didactic principles (direct instruction of reading strategies, teacher and student
modeling, and group work). In Session 2, in-depth information was provided about the
nature, function, importance, and application of the five central strategies and on the way
teachers could model the use of these strategies. Examples of modeling were provided by
means of video clips and lesson protocols. In Session 3 the focus was on reciprocal teaching
and how, by means of scaffolded instruction, the use of reading strategies is transferred to the
students. Attention was given to how the teacher can give feedback to groups of students and
how his or her expert role is gradually faded. Two training sessions for the teachers occurred
after the intervention had started, to give the teachers room to discuss their findings so far and
Teachers were given a template for the lessons that would help them keeping focused
_____________________________________________________________________
RECIPROCAL TEACHING AND TEACHER SKILLS
_____________________________________________________________________
minutes on the same day. During the classroom observations, trainers used an observation
scheme comparable to the one used by the researchers (see Classroom variables and
treatment fidelity), directing the trainers’ attention and, consequently, their feedback to the
central principles of the intervention (direct instruction of reading strategies, teacher and
Control classes
Control classes were “business as usual”. Teachers in the control classes used the regular
textbook for Dutch language that was used in their school. Among our schools, three different
language textbooks were used. The textbooks and their teacher manuals were analyzed
according to the three central principles of instructional strategies in the treatment condition:
instruction of reading strategies, modeling, and group work. Attention was given to reading
strategies in all three textbooks. However, not all strategies that were covered in the treatment
condition were also covered in the control textbooks. Reading strategies that were often
referred to were: predicting, clarifying, and attention to cohesive ties. Self-questioning did
the control classes. Almost all of the assignments were individual and there were only a few
Measures
Reading comprehension.
Reading comprehension was measured by means of the SALT-reading, a test that was
validated for use among low achieving adolescents (Van Steensel, Oostdam, & Van
Gelderen, 2013). The SALT-reading comprises eight tasks, each consisting of one or two
texts and comprehension questions about those texts. The texts cover different genres
(narrative, expository, argumentative, and instructive). They were selected from media
students assumedly come across regularly in their daily lives: (school) books, newspapers,
magazines, and official documents (such as regulations in a youth hostel). The eight tasks
comprised a total of 59 test items, that were divided into three categories: items requiring
students to retrieve relevant details from the text, items requiring students to make inferences
on a local level (e.g., draw cause-effect relationships between sentences), and items requiring
students to show their understanding of the macrostructure of the text (e.g., by inferring the
main idea of the text or the intention of the author). The test consisted mainly of multiple
choice questions but contained also five open-ended questions. The Cronbach’s alpha
coefficients of the pretest and posttest were .82 and .83, respectively.
Vocabulary knowledge.
Vocabulary knowledge was assessed with a 73-item multiple-choice test, measuring the
knowledge of nouns, verbs, adjectives, and adverbs belonging to the 23,000 words in a
dictionary for junior high school students (see Hazenberg & Hulstijn, 1996, for details). Each
item consists of a neutral carrier sentence with a bold-faced target word and four answer
options, one of which represents a correct synonym. The Cronbach’s alpha coefficient for the
IQ.
Intellectual ability was measured by administering the Raven Progressive Matrices. The total
test consists of 60 items, divided into 5 sets of 12 items. Each item represents a logical
reasoning puzzle. The items become more difficult within a set and the sets become
increasingly difficult as well (Raven, Raven & Court, 1998). For students from the lowest
tracks of prevocational education the last set was assumed to be too difficult and for this
reason this set was omitted. The Cronbach’s alpha coefficient was .82.
To examine the moderator variables, we conducted classroom observations twice during the
year. We devised an observation scheme for use both in the experimental and control
conditions. Our aim was to examine a) whether the treatment teachers gave the lessons in the
way we instructed them during the training and coaching program and b) whether the control
teachers applied treatment principles, even though they were not trained by us. The scheme
focused on three variables that were essential to the treatment: direct instruction of reading
strategies, teacher and student modeling, and group work. We constructed these variables in
the following manner, resulting in three four-point scales (0-3) to be used for further analysis:
b. Teachers introduced the central strategy of the lesson (in the treatment
condition) or any strategy (in the control condition), but provided no further
explanation (1 point).
importance, and/or application and interacted with the class about the strategy
(3 points).
behavior:
c. Teachers modeled strategy use and asked students to think aloud while using
points).
d. Teachers modeled strategy use, asked students to think aloud, and provided
b. Teachers had students work in groups, but did not provide real feedback (1
point).
c. Teachers had students work in groups and provided feedback, but not on
(3 points).
The scales were constructed in such a way that a 3-point score would be the optimal score for
the purpose of the treatment. It should be noted that the scores within a scale were
conditional: one could only score a 2 if both b and c were observed. This conditional
RECIPROCAL TEACHING AND TEACHER SKILLS
approach proved to be appropriate in the classroom observations (i.e., we did not encounter a
Before the start of the classroom observations, the observation scheme was piloted
during two lessons, one in an experimental class and one in a control class. Two researchers
filled out the observation scheme during the lessons, after which they compared their coding
and discussed causes for any differences. If these discussions revealed that items were
unclear or led to misinterpretation, the coding scheme was adjusted. Means were calculated
Procedure
The reading comprehension pretest as well as the vocabulary and IQ tests were administered
in the Fall of 2011, just before the start of the treatment, and the reading comprehension
posttest was administered during spring 2012. All test administrations took place in
classroom settings. The test sessions were introduced by a trained test leader. A familiar
teacher was present to maintain order. Questions were answered by the test leaders following
a standardized protocol.
Classroom observations took place during January-February 2012 and during April-
May 2012. During the classroom observations the researcher(s) sat at the back of the
classroom to observe the teacher. In order to be able to check codings after the observation,
There was some attrition among teachers, but not due to a lack of motivation. One
teacher in the treatment condition became terminally ill halfway during the school year.
RECIPROCAL TEACHING AND TEACHER SKILLS
Because replacement was only found after about two months, this class did not receive the
treatment in this period. When a new teacher was found, she continued giving the treatment
lessons and participated in our training program. Because of the replacement, we were not
able to do classroom observations in this class. Therefore, we were not able to include this
class in the analysis. A second teacher in the treatment condition became pregnant toward the
end of the school year. During her leave, she was temporarily replaced by a new teacher, who
continued giving the lessons and took part in the training. Finally, a teacher in the control
condition found another job halfway during the school year; a new teacher immediately
replaced her.
There was some attrition among students, mainly because of transfers to different
schools (7 students) and one student was ill for a long period of time. During the school year
Analysis
Our sample had a hierarchical structure (students nested in classes, nested in schools).
Because there was significant random variability at the class level, we performed multilevel
analyses with the use of MLwiN 2.16 (Rasbash, Steele, Browne, & Goldstein, 2009).We
tested whether a) the treatment had a significant positive effect on reading comprehension
and b) whether the quality of teacher instruction moderated the effect of the treatment.
Adding variables was done in the following order (Hox, 2010). First, all control variables
with the final three variables centered around the grand mean (Hox, 2010; Snijders & Bosker,
1999). Second, the moderator variables (strategy instruction, modeling, group work) were
entered. It was not necessary to center the moderator variables around the grand mean, as
zero was meaningful in the scoring of the classroom variables (See Classroom variables and
RECIPROCAL TEACHING AND TEACHER SKILLS
treatment fidelity) Third, the independent variable (treatment vs control) was entered to
answer the first research question. Finally, the interactions between the independent and
Of the 369 cases, 75 were incomplete due to missing values, either with missing
values within a test or questionnaire or because students were not present at one of the test
sessions due to illness (despite the fact that at each school at least one extra test session was
organized). To prevent loss of information, single imputations using SPSS missing value
analysis were performed for each variable at the item level, that is, missing items (as opposed
to ‘wrong’ items) within a test or questionnaire were imputed. No missing values were
imputed if the student was not present during the test session. As a result, 44 of the 75 cases
(58.6%) with missing values could be included in the analyses (total N = 338).
Results
Descriptive statistics
Table 1 shows the mean student scores for all continuous variables, as well as correlations
No significant differences are found between the treatment and the control condition. The
highest correlation is found between posttest and pretest reading comprehension (r=.69,
p<.01).
___________________________________________________________________
_____________________________________________________________________
In Table 2, means and standard deviations are presented for the variables resulting from the
classroom observations. As expected, the mean scores of the treatment group are higher than
RECIPROCAL TEACHING AND TEACHER SKILLS
those of the control group, indicating that in the experimental classrooms modeling, strategy
instruction and group work were more often observed than in the control classrooms. The
difference between both groups is statistically significant on the .05 level for all variables,
except for modeling. Given the small sample (N=19), the non-significant difference in the
case of modeling should not be given much weight. Bar charts (See Figures 4-6) for each
classroom variable show that only in a few instances treatment teachers scored maximally.
______________________________________________________________________
______________________________________________________________________
Multilevel analyses
As a first step, we examined whether the data had a multilevel structure. A model with only a
student level had an IGLS of 2373.309. A model with both a student and a class level had a
significantly better fit (IGLS = 2321.862; ΔIGLS = 51.447, df = 1, p < .001). Adding a school
level did not improve model fit (IGLS = 2320.221; ΔIGLS = 1.641, df = 1, p > .05).
Therefore, in all further analyses a two-level structure was used. The empty two-level model
significantly increased model fit. As expected, both reading comprehension and vocabulary at
pretest positively contributed to posttest reading comprehension. There was also an effect of
gender: boys scored significantly lower on posttest reading comprehension than girls. The
effect of IQ was nonsignificant, however: it appeared that pretest reading comprehension and
vocabulary already accounted for the variance in IQ. Therefore, IQ was dropped from the
RECIPROCAL TEACHING AND TEACHER SKILLS
model. The resulting model (Model 1, see Table 3) represented a significant increase in fit
In Model 2 (see Table 3), the moderator variables were entered. This did not result in
a significant increase in model fit (ΔIGLS = 3.095, df = 3, p > .05). This means that, overall,
Adding the treatment variable to the model (Model 3) did not result in a better fitting
model either (ΔIGLS = 0.458, df = 1 , p > .05). In other words, no main effect of the treatment
In the three subsequent models (Model 4a-c), we added the interactions between the
independent and moderator variables (i.e., the interactions of treatment and each of the three
observed instruction variables: strategy instruction, modeling and group work). Of these three
models, only Model 4a resulted in a significant improvement of fit compared with Model 3
(ΔIGLS = 5.033, df = 1, p < .05), implying that the strategy instruction variable was a
1.311, df = 14, p < .05). In other words, elaborate strategy instruction had a positive effect on
reading comprehension in the treatment classes, but not in the control classes. The size of this
effect was considerable: the interaction effect was responsible for explaining an additional 37
_______________________________________________________________
_______________________________________________________________
Discussion
Our study set out to analyze how reciprocal teaching can improve low achieving adolescents’
reading comprehension in natural classroom settings and to what extent intervention effects
are dependent on teacher behavior. Apart from analyzing the overall effects of the treatment
RECIPROCAL TEACHING AND TEACHER SKILLS
in a natural classroom setting (Research question 1), our aim was to examine whether effects
were larger when teachers provided more elaborate instruction of reading strategies, engaged
more in teacher modeling and promoted more student modeling, and when they supported
more collaboration during group work (Research question 2). Answering our first research
question, our study revealed no overall treatment effects: no significant differences were
found between students in the treatment classes and the control classes on the reading
comprehension posttest. Answering our second research question, we did find a moderator
effect of instruction of reading strategies. This moderator effect implied that in the
experimental condition more elaborate explanations of the nature, function, importance, and
The effect was substantial: it explained an additional 37 per cent of the differences between
classes after individual and class-level variables had been taken into account. In the control
condition, there was no effect of strategy instruction. It thus seems that the frequent,
systematic, and cyclical offering of reading strategies in our treatment set the stage for
there are important differences in implementation quality among teachers who are trained “on
the job”. As we have shown, neglecting such variation can result in overlooking meaningful
effects. The effect of our reciprocal teaching intervention only appeared after taking the
gives insight in the degree to which treatments such as these are successfully implemented by
‘real teachers’ (as opposed to researchers) and whether some elements are harder to apply
than others. Our experience showed that even after a year of intensive training and coaching,
RECIPROCAL TEACHING AND TEACHER SKILLS
application of the three instructional principles was less than optimal. Particularly, our
observations showed that getting students to model reading strategies during group work was
a challenge. Similar observations were made by Hacker and Tenent (2002) in an elaborate
implementation study of reciprocal teaching: they showed that teachers found it particularly
while some teachers fairly quickly succeeded in modeling reading strategies and having their
students work in groups, others had more difficulties in incorporating these principles in their
lessons. The latter seemed to be partly the result of classroom management issues: in
instances where students were unmotivated and showed oppositional behavior, teachers
of modeling or group work. There are at least two explanations for this observation. First, the
difference between instruction of reading strategies and modeling or group work can be
explained in terms of the extent of teacher versus student control. A higher score on the
strategy instruction variable indicates more elaborate instruction by the teacher about the
nature, function, importance, and/or application of reading strategies. This is the component
of the treatment that is the most teacher-controlled and is also the most familiar, both for
teachers and students (such instruction is commonly used in education, in every domain) and,
thus, is probably easiest to implement. Also, it may be assumed that teachers have prior
knowledge about reading strategies. Both direct instruction and knowledge of reading
strategies are consistent with existing knowledge and practice of teachers, whereas modeling
and group work are relatively unknown areas and therefore harder to master.
Second, modeling and group work are dependent on initiatives afforded to students:
higher scores on the former imply that more modeling is being done by both teachers and
RECIPROCAL TEACHING AND TEACHER SKILLS
students, higher scores on the latter imply more attention to group work. These components
are dependent on teachers transferring control to their students and may not be part of many
teachers’ repertoire. For low achieving students in Dutch secondary education, it is quite
uncommon that students work on tasks collaboratively in language arts lessons (De Milliano,
2013). Thus, both modeling (especially by students) and group work differ from regular
classroom practice and require new skills from teachers, as they need to adapt their feedback
to the level of the student and use techniques for motivating students to collaborate without
Therefore, one explanation of the absence of moderator effects of modeling and group
work is that these instructional strategies did not reach a certain ‘threshold level’ to become
significant moderators. The observational data of the classrooms seems to support this. The
maximum score for modeling includes students being successfully encouraged to model
reading strategy use themselves. However, only one treatment teacher managed to reach this
stage. The same holds for group work: the maximum score for group work includes teachers
changing the focus from correct responses to assignments to learning to apply reading
strategies collaboratively. Only three treatment teachers reached this stage implying that most
teachers did not attain this level of practice. Our findings are supported to some extent by the
outcomes of the study by Hacker and Tenant (2002) mentioned earlier. In this study a number
of teachers were followed over the course of one or more years to examine how they
implemented reciprocal teaching in their classrooms and to what extent they modified the
method. The authors first of all observed that “the most pervasive problem that teachers faced
with RT [Reciprocal Teaching] was getting students to learn and use the RT strategies in
group dialogues” (2002: 712). In response, teachers tended to become more directive,
providing more scaffolding in the form of whole-class instruction, teacher modeling, and
direct guidance. This was particularly true in classes with many struggling readers.
RECIPROCAL TEACHING AND TEACHER SKILLS
It appears that the training and coaching offered in our study was sufficient for the
application of one of the main elements of reciprocal teaching to bear fruit, namely strategy
instruction. However, even for this component, not all trained teachers profited sufficiently to
produce a significant difference between the experimental and the control condition. Some
experimental classes were receiving significantly more strategy instruction than others
resulting in a moderating effect of this instructional variable. For the two other main
instructional components of reciprocal teaching, modeling and group work, however, we did
not find significant moderating effects. Despite the fact that our teachers were provided with
a quite extensive training and coaching program, we believe that even more training and
coaching is needed for teachers to adapt new ways of teaching to such an extent that it
enhances the learning process of their students, as compared with control students. This is in
line with the findings of Hacker and Tenent (2002). In their research, teachers found it
difficult to embrace new practices, such as letting students work together, and clung more
As for limitations, even though there was randomization at the class and the teacher
level, students were not randomly distributed across the intervention and control group. For
randomization at the student level. Secondly, more classroom observations could have
provided more insight into the development of teachers’ implementation of the treatment, and
consequently provide more valid conclusions. In this case, two observations were enough to
find significant effects, but it would be a great addition in future research to show the
Finally, in future research we recommend a study with teachers that are trained more
extensively and more frequently than in this study. Teachers should probably be provided
with more tools to be able to guide the students in their collaborative learning process.
RECIPROCAL TEACHING AND TEACHER SKILLS
Coaching should be directed at increasing the quality of the dialogues among the students.
Teachers in our study did not have many tools to facilitate the students in their collaborative
group work. For teachers to become seasoned in new ways of teaching, they need to practice
rigorously, up to the point where reciprocal teaching becomes routine; similar to the way we
References
Braze, D., Tabor, W., Shankweiler, D.P., & Mencl, W.E. (2007). Speaking up for vocabulary:
Reading skill differences in young adults. Journal of Learning Disabilities, 40, 226
comprehension”].
http://www.nifl.gov/partnershipforreading/adolescent/default.html
of Amsterdam.
Dole, J. A., Duffy, G. G., Roehler, L. R., & Pearson, P. D. (1991). Moving from the old to the
Duffy, G. G. (1993b). Teachers’ progress toward becoming expert strategy teachers. The
an inspectorate study into language skills in the first two years of prevocationa
Edmonds, M.S., Vaughn, S., Wexler, J., Reutebuch, C., Cable, A., Klingler Tackett, K., &
Gille, E., Loijens C., Noijons, J., & Zwitser, R. (2010). Resultaten PISA-2009, Praktische
Guthrie, J.T., & Wigfield, A. (2000). Engagement and motivation in reading. In M.L. Kamil,
P.B. Mosenthal, P.D. Pearson, & R. Barr (Eds.), Handbook of reading research: Volume
Hox, J. J. (2010). Multilevel analysis: Techniques and applications. Second Edition. New
York: Routledge.
Just, M. A., & Carpenter, P. A. (1976). Eye fixations and cognitive processes. Cognitive
Just, M. A., & Carpenter, P. A. (2004). A theory of reading: From eye fixations to
Reading Association.
Kelly, M., Moore, D.W., & Tuck, B.F. (2001). Reciprocal teaching in a regular primary
LaBerge, D., & Samuels, S. J. (1974). Toward a theory of automatic information processing
Logan, S., & Johnston, R. (2009). Gender differences in reading ability and attitudes:
Examining where the differences lie. Journal of Research in Reading, 32, 199-214.
doi: 10.1111/j.1467-9817.2008.01389.x
Ministry of Education, Culture, & Science (2006). The education system in the Netherlands
2006. The Hague: Ministry of Education, Culture, & Science/Dutch Eurydice Unit.
the scientific research literature on reading and its implications for reading instruction.
Organisation for Economic Co-operation and Development [OECD] (2003). The PISA 2003
student performance in reading”, in PISA 2012 Results: What students know and can
Ouellette, G., & Beers, A. (2010). A not-so-simple view of reading: how oral vocabulary and
visual-word recognition complicate the story. Reading and Writing, 23, 189-208. doi:
10.1007/s11145-008-9159-1
Palincsar, A. S., Brown, A., & Martin, S.M. (1987). Peer interaction in reading
Pressley, M., & Afflerbach, P. (1995). Verbal protocols of reading: The nature of
RAND Reading Study Group (2002). Reading for understanding: Toward an R&D program
Rasbash, J., Steele, F., Browne, W. J., & Goldstein, H. (2009). A user’s guide to MlwiN.
Raven, J., Raven, J.C., & Court, J.H. (1998). Manual for Raven’s Progressive Matrices and
Assessment.
Rosenshine, B., & Meister, C. (1994). Reciprocal teaching: A review of the research. Review
Unrau (Eds.), Theoretical models and processes of reading. Fifth edition (pp. ).
Shadish, W.R., Cook, T.D., & Campbell, D.T. (2002). Experimental and quasi-experimental
Association.
Schiefele, U. (1999). Interest and learning from text. Scientific Studies of Reading, 3, 257–
279.
Seymoor, J.R. & Osana, H.P. (2003). Reciprocal teaching procedures and principles: two
Snijders, T. A. B., & Bosker, R. J. (1999). Multilevel analysis: An introduction to basic and
Van Gelderen, A., Schoonen, R., de Glopper, K., Hulstijn, J., Simis, A., Snellings, P. &
Van Gelderen, A., Schoonen, R., Stoel, R. D., De Glopper, K., & Hulstijn, J. (2007).
Van Steensel, R., Oostdam, R., & Van Gelderen, A. (2013). Assessing reading
Veenman, M.V.J., Van Hout‐Wolters, B.H.A.M. & Afflerbach, P. (2006). Metacognition and
1, 3‐14.
Verhoeven, L., & Van Leeuwe (2008). Prediction of the development of reading
doi: 10.1002/acp.1414
Figures
September
November
December
February
October
January
March
April
June
May
Training and Coaching Training 1 Training 2 Training 3 Coaching 1 Coaching 2 Coaching 3
Treatment Treatment in experimental classrooms
Observation Class Observation 1 Class observation 2
Students Data collection 1 Data collection 2
Strategy Example
Predicting This text has five subheadings. Write down for each subheading a)
which thoughts it evokes and b) what you already know about the
Summarizing Read the text. Read paragraph by paragraph and underline in each
one or two sentences summarizing it. Use the words you underlined.
Self-questioning Read the text. Note at least five questions that spring to mind while
reading.
Clarifying Search the text for difficult words. Try to uncover their meaning using
these hints: a) reread the previous piece of text or read on, b) look at
the illustrations in the text, c) look at the word: you might know part of
the word, d) sometimes you have to use your own knowledge to figure
Interpreting Read the text. Underline the signal words. Answer the questions, while
however]
hence]
Write the subject of the text and the central strategy of the lesson on
the blackboard.
Introduction Write down questions students have about the text during orientation.
Read the first paragraph together and model the central strategy.
Invite a student to read the next paragraph while thinking aloud and
applying the central strategy. Give support when necessary, that is,
Appendix).
Processing
Walk around to give the groups of students feedback. Focus on the
central strategy and motivate the students to apply the strategy while
content.
Reflection
Together with the students, answer the questions they had before
reading the text. Did reading the text answer those questions?
0
0 0,5 1 0 0,5 1 1,5 3
Control Treatment
Figure 4. Bar chart for modeling, for both the control and treatment teachers.
0
0 0,5 1,5 1 1,5 2 2,5 3
Control Treatment
Figure 5. Bar chart for strategy-instruction, for both the control and treatment teachers.
RECIPROCAL TEACHING AND TEACHER SKILLS
0
0 1 1,5 0 1,5 2 2,5 3
Control Treatment
Figure 6. Bar chart for group work, for both the control and treatment teachers.
RECIPROCAL TEACHING AND TEACHER SKILLS
Tables
Table 1
1. Posttest reading 37.37 (7.20) 36.22 (8.91) 1.30 .69* .54* .25*
comprehension
comprehension
*p<.01
Table 2
Note. Scoring between the three variables cannot be compared one-on-one. The meaning of
the scoring (0-3) is different for each variable. See Classroom variables and treatment fidelity
Table 3
Intercept 36.409 (0.984) 37.431 (0.625) 37.398 (0.857) 37.875 (0.912) 36.621 (0.950) 37.460 (0.975) 38.637 (1.245)
Gender (Male = 1) -1.442 (0.600)* -1.364 (0.601)* -1.368 (0.601)* -1.363 (0.600)* -1.355 (0.601)* -1.370 (0.601)*
Reading comprehension pre 0.572 (0.048)*** 0.570 (0.048)*** 0.572 (0.048)*** 0.572 (0.048)*** 0.581 (0.049)*** 0.573 (0.048)***
Vocabulary pre 0.167 (0.043)*** 0.166 (0.043)*** 0.166 (0.042)*** 0.171 (0.042)*** 0.167 (0.042)*** 0.165 (0.042)***
Strategy instruction 1.144 (0.704) 1.327 (0.746) -0.201 (0.898) 1.379 (0.724) 1.588 (0.787)
Modeling -0.048 (0.883) 0.004 (0.876) -0.171 (0.766) -1.380 (1.616) 0.018 (0.857)
Group work -0.252 (0.510) -0.077 (0.565) -0.618 (0.540) -0.161 (0.554) 0.803 (1.137)
Treatment (treatment = 1) -0.976 (1.434) -0.210 (1.276) -0.680 (1.418) -1.650 (1.595)
Treatment * strategy instruction 3.183 (1.311)*
Treatment * modeling 1.866 (1.854)
Treatment * group work -1.211 (1.371)
Level 1 random variance 50.831(4.025) 28.138(2.228) 28.129 (2.227) 28.129 (2.227) 28.160 (2.229) 28.150 (2.229) 28.139 (2.228)
Level 2 random variance 15.379(5.961) 4.267(1.920) 3.404 (1.638) 3.285(1.599) 2.068 (1.198) 2.979 (1.497) 3.063 (1.526)
Level 1 explained variance 45% 0% 0% 0% 0% 0%
Df 3 3 1 1 1 1